This workshop emphasized small-group discussion as well as addition to speaker presentations to gain a greater understanding of the key issues regarding the measurement of costs and benefits of the food system through in-depth expert discussions on focused topics. The complete second half of the first day was spent with participants divided into working groups, with each working group focused on one of four categories of health or environmental effects: (1) energy and greenhouse gas (GHG) emissions; (2) soil, water, and other environmental consequences; (3) health consequences of antimicrobial use in agriculture; and (4) other public health consequences. Although the working groups focused distinctly on these categories, several participants highlighted the underlying complex linkages and interactions between them. For example, concentrated animal feeding operations may provide improved control of some pathogens of public health concern and create less waste per pound of product, but these large operations also produce considerable quantities of manure that may lead to run-off problems when applied to nearby cropland. To reiterate the linkages and provide a more complete picture, some effects may appear in more than one of the subsequent working group summaries.
Each working group was given a matrix worksheet (Table 6-1) that focused on the following six key items for consideration:
- The source(s) of the effect
- Whether the effect is an environmental, public health, or other type of effect
- Methodologies and limitations to measuring the effect
TABLE 6-1 Explanation of the Matrix Provided to Working Groups to Report on Effects of Practices Spanning the Life Cycle of Foods
|Source(s) of Effect||Environmental Effect||Measurement/Limitations||Externality?||Public Health or Other Effect|
|What practice or action is the primary source of the effect?||Describe the effects of practices that have an impact (positive or negative) on the environment.||What indicators or methods can be used to measure this effect?||Is the effect an externality?||Describe the effects of practices that have a public health, economic, or social impact (positive or negative).|
|For example, are the effects from crop production, manure management, fertilizer use, from use of farm machinery, transportation, or dietary intake?||What methodological limitations may inhibit measuring this effect?||Yes (Y), No (N), or Unclear (U)||If this effect is an indirect or mediated effect of an environmental effect, list the related effects in the same row. If this effect is not indirectly linked to an environmental effect, start a new row.|
- Whether the effect is an externality
- Trade-offs related to the source of the effect and methodologies and limitations to measuring trade-offs
- The life cycle stage during which the effect occurs (e.g., production, processing, distribution)
The groups took different approaches to addressing these items, with some participants noting that the matrix did not allow for an exhaustive examination since components, like the magnitude of the effect, are missing. This chapter summarizes the discussions that took place during the small-group discussions.1
1 The summaries of the working group discussions are intended to demonstrate the diversity of perspectives and divergent opinions and should not be construed as reflecting any group consensus.
|Measurement/Limitations||Externality?||Trade-offs Related to Alternative Strategies||Measurement/Limitations||Life Cycle Stage|
|What indicators or methods can be used to measure this effect?||Is the effect an externality?||What are the trade-offs (economic, environmental, health, or other) associated with the practice(s) that are causing the effect?||What indicators or method can be used to measure the trade-offs?||At what stage of the life cycle is the source of this effect occurring? Choose one of the following:|
|What methodological limitations may inhibit measuring this effect?||Yes (Y), No (N), or Unclear (U)||What methodological limitations may inhibit measuring the trade-offs?||
• resource origin
• agricultural production
• food processing, packaging & distribution
• preparation & consumption
Key points from the discussion are summarized here, as reported back to the group at large by Greg Keoleian, the Peter M. Wege Endowed Professor of Sustainable Systems at the University of Michigan. According to Keoleian, the group observed that the matrix (Table 6-1) could be useful, but decided that filling it out would have been too time consuming, given how much is already known about GHG emissions and energy use.
LCA was perceived as the tool of choice for evaluating both GHG emissions and energy use in a comprehensive way by individual participants of the group. With respect to GHG emissions, LCA could be used to evaluate both major emissions (i.e., CO2, CH4, N2O) and minor emissions (e.g., chlorofluorocarbons used as refrigerants; perfluorocarbons used during aluminum production). Some emissions data are available at the national level (e.g., Environmental Protection Agency [EPA] reports on
emissions associated with various agricultural activities, such as CH4 emissions related to rice cultivation). With respect to energy use, LCA could be used to account for all of the various energy carriers across the total fuel cycle (e.g., transport fuels, electricity), as well as for upstream energy sources (e.g., feedstock for agricultural chemicals and fertilizers). With respect to which analytical approach to take, one could conduct either an attributional or consequential LCA, depending on the research question. (See the summary of Marty Heller’s presentation in Chapter 3 for descriptions of the two approaches.)
Defining System Boundaries and Unit of Analysis
While LCA extends across all stages, from feed to end-of-life (i.e., feed, farm operations, processing, retail, consumption, disposal or end-of-life), the analysis could be truncated so that only certain components are evaluated. For example, in a comparison of agricultural production methods, it would not be necessary to include product packaging. According to Keoleian, the group spent a great deal of time discussing the importance of defining the functional unit, that is, the basis of analysis (e.g., kilograms [kg] of meat, kg of protein, calories, total nutrition or diet), and the importance of defining the temporal and spatial boundaries of the analysis.
Effects to Consider
Many GHG emission and energy use effects are quantifiable. The “real issue,” Keoleian reported, is uncertainty. Many effects are difficult to accurately estimate. Keoleian described CO2 from fuel combustion, N2O from soils and manure, and CH4 from manure and enteric fermentation, all of which contribute to climate change, as important direct costs to consider; and CO2 sequestration resulting from certain types of land use changes (e.g., converting marginal land to rangeland) as a potential benefit to consider. Indirect land use change impacts can be important, but are difficult to quantify.
With respect to energy usage, Keoleian reported that British thermal units of primary energy consumption can be quantified “pretty well.” LCA can also be used to quantify impacts from air pollutant emissions associated with energy usage (e.g., NOx, PM, Hg, SO2 from coal combustion); water pollution (e.g., nitrate run-off from corn production, oil spills); and land use impacts such as biodiversity loss (e.g., from surface mining of coal).
Multiple working group participants emphasized several measurement challenges:
- A high degree of uncertainty in characterizing non-CO2 emissions—Several different chamber and field measurement methods could be used to estimate the emissions. Using Intergovernmental Panel on Climate Change factors for ruminant enteric fermentation emissions, for example, is not an accurate method to estimate these emissions. According to Keoleian, many working group participants emphasized the development of new biogeochemical models that help to better characterize emissions.
- Heterogeneity in production methods—Different production methods can have different impacts, yet LCA data tend to offer limited resolution of these differences, calling for more extensive research exploiting the heterogeneity.
- Non-GHG air pollutant emissions have both regional and local effects—The impacts related to emissions of SO2, NOx, mercury, and other pollutants are more site dependent than GHG emissions, so it is important for the location of the emissions to be inventoried. Unfortunately, many databases do not report emissions in a spatially explicit manner. On the other hand, for carbon emissions and climate change impact, it does not matter where the greenhouse gases are released.
- Allocation rules can influence results—Allocation rules are used to distribute impacts from processes with coproducts across the various outputs (e.g., allocating feed production burdens to milk, butter, and hides).
The working group participants discussed several activities that drive emissions and energy use: feed production, enteric fermentation, manure management, food storage (i.e., refrigeration), and food waste. For example, an estimated 26 percent of edible food is wasted. Participants highlighted several potential improvement strategies for countering these effects: adjusting animal rations and managing feed quality; harvesting energy from manure through anaerobic digestion; substituting renewable energy sources; and shopping more frequently to reduce household refrigeration.
Trade-Offs Related to Alternative Strategies
Many participants of the working group recognized that, when considering GHG emissions and energy usage, particularly when considering policies and interventions aimed at reducing GHG emissions or energy usage, one must also consider the human health, environmental health, and economic trade-offs. For example, with respect to the size of a production operation, while some concentrated animal feeding operations (CAFOs)
may be more efficient than smaller operations, there may be trade-offs with respect to water quality, manure management, and other effects.
The LCA framework is very good at characterizing the effects of both production and consumption, particularly with respect to GHG emissions and energy usage. While the framework can be used to also characterize human health (e.g., via quality-adjusted life years) and other social impacts of production and consumption, Keoleian reported, “uncertainty increases tremendously when you start to look at some of these other effects.” However, the LCA framework can be very useful in identifying “order of magnitude” trade-offs between health and environmental impacts. LCA can also be used to evaluate economic impacts of production and consumption, including both private and social costs (e.g., the “social cost of carbon,” that is, monetized damages associated with increasing carbon emission) (Interagency Working Group on Social Cost of Carbon, 2010).
Research and Data Needs
Data needs depend on the question(s) being addressed. For example, if a goal is to characterize differences in production methods, then data would be needed for each type of production method (i.e., as opposed to industry average). With respect to data needs for specific stages of the food life cycle, many working group participants indicated there could be better data on the generation of food waste (e.g., data on spoiled milk is decades old) and better data on consumption patterns. Spatially explicit production data will also be necessary to capture impacts of categories that have spatially influenced characterization factors (e.g., water use, eutrophication, land use).
Participants in this group spent most of their time discussing challenges to characterizing the soil, water, and other environmental consequences of the food system, as reported back to the group at large by Justin Derner, research leader for the Rangeland Resources Research Unit of the USDA Agricultural Research Service.
The Challenge of Heterogeneity
Working group participants discussed several major challenges to analyzing the external costs of animal production. One main challenge is the heterogeneity among sites with respect to practices, soils, climate, landscape, plant communities, and data (e.g., some sites have plentiful data, others none). Also, effects occur across variable spatial scales (e.g.,
small-scale farms versus large rangelands that encompass hundreds of thousands of hectares and may be publicly managed) and temporal scales (e.g., short-term versus long-term effects). On top of all this already existing heterogeneity, climate is not only changing, but it is changing differentially across the landscape, and the human population is growing, creating new food demands.
Building a Framework
It was suggested that one way to build a framework for addressing the environmental costs of the food system is to consider the threshold or cut-off rates of application beyond which four key elements—carbon (C), nitrogen (N), phosphorous (P), and sulfur (S)—become pollutants instead of nutrients. The analysis would very site-specific, but at least it would provide a framework for moving forward.
Effects to Consider
Although they did not identify externalities in the pure economic sense of the word, the group participants considered a wide range of effects: soil water erosion, soil wind erosion, soil fertility, water quality, water quantity, water scarcity, biodiversity, air quality/odors, pesticides, herbicides, open spaces, genetically modified organisms (plant and possibly animal), land use change, and deforestation. Additionally, there are several fairly well-known public health effects to consider in relation to some of these environmental effects, for example, asthma and mental health effects associated with exposure to certain odors. Several other considerations not captured in terms of monetization came up during conversation: quality of life; connection to the land; the value of open and green space; animal welfare issues; salt accumulation in soils; the value of wildlife habitat; ecosystem resilience (i.e., some ecosystems are resilient even after abuse, and show no change even when “pushed to the limits,” while others are more fragile and undergo dramatic changes); and weed resistance to herbicides (e.g., some herbicides induce dramatic changes in ecosystem production).
Sources of Information and Challenges in Analyzing the Data
With respect to data, plentiful data are already available in various data networks and databases. Derner mentioned the Long-Term Ecological Research (LTER) Network and the new National Ecological Observatory Network (NEON), both funded by the National Science Foundation (NSF); the Long-Term Agro-Ecosystem Research (LTAR) network and Greenhouse gas Reduction through Agricultural Carbon Enhancement network
(GRACEnet), both coordinated by the U.S. Department of Agriculture’s Agricultural Research Service (ARS); AmeriFlux; and citizen science efforts (i.e., public participation in scientific research). In addition to these data-collecting networks, several existing databases could be useful, such as the Conservation Effects Assessment Project (CEAP) database, long-term data at many sites, meta-analyses data, and remote sensing data.
While the data may be plentiful, so too are the limitations to analyzing those data. For example, the group struggled with identifying a benchmark for analysis. That is, what qualifies as “conventional” practice?2 “Conventional” practices evolve over time. Additional challenges include the detection of “improper” management; the spatial distribution of manure/urine from animals; legacy effects of the dust bowl (e.g., huge soil losses); a realization that the global supply of phosphorous is limited and predicted to be depleted in less than a century, with consequences for cropping systems; the likelihood that there may be “sensitive areas” of high concern that could be targeted for sampling, with a cluster analysis focused on those areas; and water laws/rights and their impact on the cost of food.
This group focused most of its discussion on swine production, reported facilitator Michael Doyle, Regents Professor of Food Microbiology and director of the Center for Food Safety at the University of Georgia.
State of the Evidence
Scientists have used a variety of tools—epidemiology, risk assessment, and molecular biology—to collect evidence on the public health impact of the use of antimicrobials in food production. Much of the epidemiological evidence resides with the U.S. Centers for Disease Control and Prevention (CDC). Most of the evidence is indirect findings related to antimicrobial use in agriculture.
Key challenges to collecting even indirect evidence, but especially direct evidence, stem from the complexity of the emergence of antibiotic resistance. This resistance can emerge in any of several ways: transfer between species, acquisition from the environment, selection, or co-selection. Co-selection occurs when use of an antibiotic selects not only for a resistance gene against the antibiotic being used, but also for resistance genes against other antibiotics. “We need to learn considerably more about co-selection,” Doyle reported.
When thinking about the public health impact of antibiotic resistance,
2 Some workshop attendees disapproved of the use of the word “conventional.”
it is important to consider not just increased morbidity and mortality, including the potential for untreatable disease (e.g., systemic Salmonella infection that would be untreatable with antibiotics), but also the fact that antibiotic use in food production creates an environmental reservoir of antibiotic-resistant genes that includes non-pathogenic bacteria (i.e., nonpathogenic bacteria can harbor antibiotic-resistant genes that can be transferred to pathogenic bacteria).
There was disagreement about the degree of evidence needed to establish a relationship between the use of antimicrobials in animal food production and human health and the feasibility of obtaining risk assessment data. Some individuals suggested that risk assessments are necessary and have been used successfully in the past to address this issue. Others expressed concern that risk assessments modified to evaluate risk from the use of antimicrobials in agriculture would be too costly and that conducting risk assessments on every antibiotic in every animal species would not be feasible.
Trade-Offs Related to Alternative Strategies
Working group participants highlighted several trade-offs that could be considered when evaluating the effects of antibiotic use in food production. First is productivity, with antibiotic use resulting in a more rapid growth rate and increased productivity, which in turn can reduce production costs (e.g., a more rapid growth rate can result in less manure and thereby lower the cost associated with removing manure). A second trade-off to consider is animal health and welfare. Third is food safety, with slower growth rates sometimes being associated with increased prevalence of disease. In high-intensity poultry production (i.e., with the use of antibiotics), the average time to grow a chicken from 1 day to age of processing is 42 days. Without the use of antibiotics, the average time increases by several days. The longer the production time, the greater the risk of Campylobacter colonization of poultry, a major cause of foodborne diarrheal illness in humans. A fourth trade-off to consider is profit, with discontinuation of the use of antibiotics being an increased cost for the farmer in animal growth rates and increased potential for disease. A final trade-off to consider is that a more efficient production process creates less waste per pound of product.
Research and Data Needs
Multiple working group participants voiced support for several research and data needs that would improve understanding of the public health ef-
fects of antimicrobial use in food production. Participants suggested more research could be conducted specifically to meet the following data needs:
- Data on the use of antimicrobials in agriculture—not just which antimicrobials are being used for which animal species, but also how they are being used with respect to dose, duration, and frequency
- Data on antibiotic use in humans (same types of data as listed above)
- Data on the evolution and transfer of resistance genes in different types of bacteria
- Data on co-selection
- Data on the prevalence of antimicrobial-resistant microbes—that is, the number of animals in an animal production facility actually carrying resistant strains of potentially harmful microbes
- Data on the impacts of different farm practices on disease management (e.g., Doyle suggested that Denmark would be a good place to start with respect to studying the impact of different farm practices, given its major strides in reducing antimicrobial use without impacting production cost or efficiency)
Many participants also emphasized possible improvements to the National Antimicrobial Resistance Monitoring System, which is the main system used to monitor antimicrobial resistance in animals, humans, and meats. While the system has been up and running for about 10 years and has revealed some trends, there are concerns that it is not well integrated (i.e., the database could be redesigned in a way that makes it easier to correlate antimicrobial resistance trends in animals, humans, and meats), and that it does not monitor emergent pathogens not traditionally found in foods (e.g., methicillin-resistant Staphylococcus aureus, Clostridium difficile, urinary tract E. coli).
Participants in this group worked with the matrix (Table 6-1), not always exhaustively filling in the matrix worksheet, rather discussing the six key items in turn. Included here is a summary of the report-back to the group at large by Sandra Hoffmann, senior economist with the Food Economics Division of the USDA Economic Research Service. The report identified 10 broad lessons drawn from the working group’s discussion.
Potential Public Health Effects
Many participants in this group recognized that the American food system provides significant health benefits, in particular the provision of
affordable nutrition, but given the task of the workshop, discussion focused on adverse health effects. Working group participants discussed a wide range of potential adverse health impacts from food production, processing, marketing, and consumption. Among these were
- acute and chronic illness from foodborne pathogens and parasites (e.g., enterohemorrhagic E. coli in beef, Salmonella in poultry or produce, and parasites like Toxoplasma gondii);
- the effects of exposure to chemicals (i.e., drug residues, hormones, and environmental toxins);
- diet-related chronic disease (e.g., diabetes, cardiovascular disease, cancer);
- occupational injuries and disease associated with agricultural production and food processing;
- adverse health effects associated with transportation (e.g., motor vehicle crashes, effects of air pollution);
- effects of exposure to air and water pollution from production practices (e.g., pesticide drift, manure-related ammonia emissions, and polluted surface water);
- mental health impacts (e.g., mental stress associated with living or working near concentrated animal feeding operations [CAFOs] or with living and working conditions among migrant laborers); and
- social impacts (e.g., effects of CAFOs on independence of rural communities, rural development, ability to conduct social or leisure activities) (see also Donham et al., 2007).
There was discussion about how much evidence of causality, as opposed to association, is necessary to identify an effect. The working group participants viewed their task for this exercise as discussing the scope of possible adverse health effects. Many participants recognized the importance of further work that would help to establish causality and to quantify the extent of the impacts. As one participant said, “This is just hypothesis generation at this point.”
Working group participants discussed the availability and usefulness of different data sources that could be used to quantify these impacts. In general, there are limits to the usefulness of disease surveillance data in providing a comprehensive picture of health patterns associated with food production and consumption. Chemical exposures in agricultural production can result in acute illness. Those poisoned may seek care and cases may be reported to public health authorities. But there can be long latency
periods between chemical exposure and illness, making it difficult to establish causation through surveillance data. Dietary exposure is typically very low-level, though potentially over long time periods. For these reasons, disease associated with chemical exposure is typically based on estimates of exposure and dose-response rather than surveillance data. Surveillance data are more useful in quantifying foodborne illness from pathogens or parasites, which are frequently associated with an acute onset of symptoms. But even then, most who suffer from these acute illnesses do not seek medical care and care providers may not report illnesses they do see. In addition, the availability of medical care varies geographically and by socioeconomic group. Active surveillance is not conducted on many of the outcomes of concern or in many areas of the country. As a broad generalization, the quality of data on exposure and dose-response modeling is generally stronger for chemicals than for foodborne pathogens and parasites, and the quality and availability of disease surveillance data are generally stronger for pathogens and parasites than for chemical hazards. Federal agencies, state departments of public health, and possibly some private-sector organizations will be important sources of data; examples include the EPA for air and water pollution exposure data, the National Institute for Occupational Safety and Health, state health departments and possibly labor unions for occupational safety data, and the CDC for surveillance data on foodborne illness and other health outcomes. The 2003 Institute of Medicine (IOM) report Dioxin and Dioxin-Like Compounds in the Food Supply: Strategies to Decrease Exposure provides data on exposure to dioxin in food (IOM, 2003).
Externality or Not?
Group participants discussed which effects could be considered externalities. There was lively debate about whether a full-scale accounting of the food system should even cover external effects that have been internalized. For example, participants debated the extent to which the cost of diet-related cardiovascular disease is internalized through health and life insurance premiums and therefore not an externality and should not be included in the analysis. Many participants agreed that even if the cost of health care were fully internalized through insurance premiums, the impact of disease extends beyond the ill person and their immediate family, and therefore social costs likely exceed medical costs. As a result, the full social cost is not internalized by individuals, and would be important to reflect in external costs. Group participants recognized that the analysis required to adequately address these kinds of questions was beyond this scoping exercise. In the spirit of viewing the discussion as an effort in hypothesis generation, participants decided to identify all potential health effects, regardless of whether those health effects qualify as externalities.
Trade-Offs Related to Alternative Strategies
The group had limited time to discuss trade-offs among alternative strategies to reduce external costs associated with food production and consumption. One set of trade-offs discussed was those related to large- versus small-scale production. Potential benefits of CAFOs include an economy of scale that affords more efficient sewage and manure management and, in some cases, improved control of some pathogens. For example, trichinosis from pork has been significantly reduced by the improved rodent control made possible by confined feeding operations. Potential costs include the mental health and community effects where CAFOs are located, and possibly greater prevalence of other pathogens or greater use of antibiotics among CAFOs compared with smaller-scale livestock operations.
Hoffmann summarized what she viewed as the 10 major lessons from the working group exercise:
- The matrix did not include some important dimensions of the problem components. In particular, it did not provide a place to include the magnitude of the impact and confidence about the magnitude of impact. It also did not provide a place to note the distribution of impacts. For example, impacts may vary by geographical location or by income, age, or social groups.
- The concept of externality might not be the best way to frame the analysis because it does not capture or allow expression of some major concerns in the public health community. For example, cardiovascular disease is a major cause of death in the United States. Yet, it is not clear that diet-related cardiovascular disease is an externality.
- Many participants of the working group felt that more consideration could be given to methods or approaches for capturing the social and individual impacts of large-scale production, for example, impacts of CAFOs on local social networks and local energy use.
- Focusing exclusively on adverse health impacts of food production and consumption without also looking at the health benefits may provide a distorted picture. For example, while excessive red-meat consumption can contribute to cardiovascular disease, meats also provide a high-quality source of protein. There was also some discussion on whether an examination of nutritional benefits should focus on individual foods or dietary patterns at large.
- The industrial structure of production is important to take into account. For example, the cost of seeking other work with fewer occupational safety risks is greater in rural areas where meat processing plants are isolated and where workers have fewer options for employment.
- Regional concentration is also important to consider. For example, toxic algal blooms have appeared in areas where there is high regionally concentrated agricultural production, such as in the Delmarva Peninsula in the Chesapeake Bay (where poultry production is concentrated) and in the Carolinas (where hog production is concentrated).
- Capturing all stages of the life cycle, not just production and consumption, would be useful. For example, the group did not discuss food preparation by the retail sector of the food industry. Yet, the public health impact associated with the addition of salts, nitrates, or other additives to foods in restaurants or other retail establishments could be considered an externality if consumers are unaware of the addition or risks of those substances.
- Considering production methods is important when evaluating health impact. Certain foodborne illnesses are reemerging in association with changes in management practices. For example, trichinosis is reemerging in association with field-raised hogs, and some dairy-related illnesses are reemerging with a loosening of norms around pasteurization.
- While the group discussion was a great brainstorming exercise, moving forward will require a very solid literature review and analysis of available data. The matrix helped facilitate group discussion, but it may not be the best structure for more in-depth analysis.
- It is important to define the scope of the effects to be considered in a full report. For example, is the goal to examine only direct public health effects of food consumption or to more broadly examine indirect effects of the production process as well (e.g., occupational illness)? Also, is the goal to examine acute effects, chronic effects, or transgenerational effects?
Although the groups took different approaches, based on the report-backs to the group at large, the group discussions shared several major overarching themes:
- Many public health and environmental costs can be quantified, but there is a great deal of uncertainty about many estimates. For example, with respect to energy and emissions, most non-GHG emission amounts are significantly uncertain.
- Production systems are highly variable, not just with respect to methods (e.g., large- versus small-scale production), but also site specificity (e.g., local soils, climate, landscape), which has implications not just for analysis but also for data collection.
- Even people with very different perspectives together voiced their support for more information that would allow for improved decision making about many of the issues. For example, many data sources could be updated, particularly with respect to consumption.
- Many questions about the scope of effects need to be considered, with varying opinions about whether the concept of externality is the best way to frame a full-scale accounting of the “true costs” of food. All four break-out groups struggled to understand exactly what to measure—externalities as defined by economists or all external effects regardless of whether they qualify as externalities (e.g., external effects that are internalized).
- All groups recognized the importance of trade-offs. Indeed, workshop chair Helen Jensen began the break-out group report-back session by commenting on the April 2012 announcement that the European Union would be banning sow stalls beginning in January 2013. The predicted 5-10 percent price increase for pork as a result of the ban is a good example of the type of trade-off that needs to be considered when evaluating the effects of different regulations or practices (European Commission, 2012).
- Jensen observed that everyone began to get a better sense of the food system and began to see problems somewhat differently during the small group discussions.
- The groups were largely brain-storming exercises. There were several calls for a more systematic approach to identifying effects and methodologies for measuring those effects. For example, one participant suggested that a systematic survey of the literature would yield more comprehensive lists of effects, trade-offs, methodologies, and limitations of those methodologies.
Donham, K. J., S. Wing, D. Osterberg, J. L. Flora, C. Hodne, K. M. Thu, and P. S. Thorne. 2007. Community health and socioeconomic issues surrounding concentrated animal feeding operations. Environmental Health Perspectives 115(2):317-320.
European Comission. 2012. Animal welfare: Commission steps up pressure on member states to implement ban on individual sow stalls. Press release. http://europa.eu/rapid/pressReleasesAction.do?reference=IP/12/404 (accessed October 19, 2012).
Interagency Working Group on Social Cost of Carbon. 2010. Technical Support Document: Social cost of carbon for regulatory impact analysis, Under Executive Order 12866. Washington, DC: Environmental Protection Agency. http://www.epa.gov/oms/climate/regulations/scc-tsd.pdf (accessed November 5, 2012).
IOM (Institute of Medicine). 2003. Dioxins and dioxin-like compounds in the food supply: Strategies to reduce exposure. Washington, DC: The National Academies Press.